Nathan F. Sayre
Recognizing History in Range Ecology:
100 Years of Science and Management on
the Santa Rita Experimental Range
Abstract: At the centennial of the Santa Rita Experimental Range, historical analysis is called
for on two levels. First, as a major site in the history of range ecology, the Santa Rita illuminates
past successes and failures in science and management and the ways in which larger social,
economic, and political factors have shaped scientific research. Second, with the turn away from
equilibrium-based models in range science—a turn prompted in part by research at the Santa
Rita—there is a growing need for history in range ecology itself. I discuss the needs, premises, and
events underlying establishment of the Santa Rita in 1903. Then I examine the evolution of
research and management recommendations through four major periods from 1901 to 1988, and
I discuss the land swap that transferred the Santa Rita to State ownership in 1988 to 1991. Finally,
I consider what effects the Santa Rita has had on rangelands and range management in the region.
I argue that a static conception of the carrying capacity of Southwestern rangelands was imposed
for economic and political reasons, over the objections or reservations of early range scientists at
the Santa Rita, and that this may have contributed both to range depletion and to rancorous
relations between public agencies and private ranchers in the twentieth century. To meet society’s
current demands on rangelands, the long-term, large-scale data assembled from the Santa Rita
will be critically important.
Keywords: range science, range ecology, history, carrying capacity, mesquite, Frederic
Clements, semiarid rangelands
Acknowledgments: The author gratefully acknowledges Joel Brown, Kris Havstad, Mitchel P.
McClaran, Guy McPherson, and especially Phil Ogden for helpful comments and other assistance
in the preparation of the manuscript. Thanks also to Barbara Gibson for help with the figures.
I remain wholly responsible for the contents.
Introduction _____________________________________________________
The Santa Rita Experimental Range is 100 years old this year, providing an occasion to celebrate and to reflect. The first
of many experimental ranges in the United States, the Santa Rita was founded at a time when both range science and plant
ecology were in their infancy. The purpose was to conduct research that would aid in the management of Southwestern
rangelands by public agencies and private ranchers, in the belief that science, coordinated by public agencies and conducted
on a suitably large scale, would produce methods of restoring and conserving the vast and severely degraded rangelands of
the region more quickly and effectively than a private, trial-and-error approach could. Confidence in the ability of government
science to solve pressing public problems was characteristic of the era, giving birth not only to the Santa Rita but also to range
science more generally and to an array of Federal agencies.
To assess a century of work on the Santa Rita, at least two questions must be answered: (1) What happened on the
experimental range itself, in terms of research and recommendations for management? And (2) what effects did this work
have on rangelands in the region? The historical record is abundant regarding the first question, but comparatively thin
as to the second. I begin by reviewing the circumstances surrounding the creation of the Santa Rita Experimental Range.
Then I use the more than 400 publications produced from the Santa Rita to define four major periods of research from 1901
Nathan F. Sayre is Postdoctoral Research Associate with the Jornada Experimental Range in Las Cruces, NM. Phone: 520-327-5567; e-mail: Nfsayre@aol.com.
He earned a B.A. degree in philosophy from Yale University and a Ph.D. degree in anthropology from the University of Chicago. His research focuses on the history
of range science and management in the Southwest, and the implications of that history for current management.
In: McClaran, Mitchel P.; Ffolliott, Peter F.; Edminster, Carleton B. tech. coords. Santa Rita Experimental Range: 100 years (1903 to 2003) of accomplishments
and contributions; conference proceedings; 2003 October 30–November 1; Tucson, AZ. Proc. RMRS-P-30. Ogden, UT: U.S. Department of Agriculture, Forest
Service, Rocky Mountain Research Station.
USDA Forest Service Proceedings RMRS-P-30. 2003
1
Sayre
to 1988. Within each period, I examine selected documents—some published, some unpublished—to trace the
evolution of research questions and management recommendations. Then I briefly discuss the period since 1988,
when ownership of the range changed from Federal to Sate.
Finally, I examine the evidence regarding actual management and range conditions over the past 100 years. Although the degree of influence of the Santa Rita is difficult
to determine in detail, several themes and possible lessons
for the future emerge nevertheless.
The overarching thesis of my argument is that a century
of research at the Santa Rita indicates the need for historical
analysis both of and in range ecology. Understanding the
history of range ecology is important for the same reasons as
in any discipline: to learn from past failures and successes,
to recognize intellectual antecedents, and to enable critical
reflection on our own ideas and practices. The history of the
Santa Rita reveals that while the methods and emphases of
research changed to reflect accumulating knowledge, the
central questions and many management recommendations
remained surprisingly consistent until very recently; it also
suggests that institutional and political factors have been as
important as scientific or ecological ones in shaping the
knowledge that researchers produce. The importance of
history in range ecology emerges from what has been learned
in the past century, both at the Santa Rita and in other arid
and semiarid settings. Whereas equilibrium-based ecological theory allowed most past researchers to neglect historical questions, current theories emphasize the potential of
nonstationary climate and discrete events, interacting at
various spatial and temporal scales, to cause significant and
lasting ecological change. Today, with the larger social,
political, and economic contexts of range management dramatically different from a century ago, there is a need both
to recognize and to re-cognize history, so that the changes of
the past can be properly understood and the challenges of
the present and future effectively confronted.
Recognizing History in Range Ecology: 100 Years of Science and Management...
contained a prescient warning. In the open range, Toumey
wrote (1890: 7):
...even the hardiest grasses when continually eaten close to
the ground will, as a rule, in a few years become extinct…
[W]here drought and overstocking both combine, and the
grass that does not burn out from the effects of the hot sun, is
continually eaten close to the ground by hungry cattle, the
range is in poor condition to produce feed for the following
season. The repetition of this process year after year cannot
help but decrease the supply of grasses on the range.
By the time the drought broke in late summer 1893, an
estimated 50 to 75 percent of southern Arizona’s cattle had
perished from lack of feed or water. Photographs from the
time bear out Toumey’s most dire scenario (fig. 1).
The undeniable ecological and economic damage of the
drought helped get the attention of Congress, which in
1895 appropriated the first Federal funding expressly for
range research. In the Texas high plains, Jared Smith and
H. L. Bentley arranged to fence two sections of rangeland
Beginnings: Founding a Range
Research Reserve ______________
Nineteen hundred and three was the fifth year of a 6-year
drought in southern Arizona. The boom and bust cycles of
markets and rainfall were already painfully familiar to both
ranchers and public officials in the area. Just 10 years
before, the drought of 1891 to 1893 had killed scores of
thousands of cattle and wiped out countless ranchers. That
drought, more than the one from 1898 to 1904, helped to set
in motion the factors that would eventuate in the Santa Rita
Experimental Range. But it was not the only, or the first,
factor.
The Hatch Act of 1887 authorized State and territorial
governments to receive Federal funding for agricultural
experiment stations. Lacking other resources, the University of Arizona took advantage of Hatch Act funds beginning in 1890, using them to cover operating expenses and
salaries as well as agricultural research (Webb 2002: 80). A
year later the Arizona Agricultural Experiment Station
published its Bulletin number 2, comprising two short
articles by J. W. Toumey (1890): “Arizona Range Grasses in
General” and “Overstocking the Range.” The latter article
2
Figure 1—Photograph taken by George Roskruge,
surveyor for the General Land Office, at an unidentified southeastern Arizona location in the
summer of 1891. Heavy, unconrolled grazing combined with drought produced widespread denudation of rangelands previously dominated by
perennial bunchgrasses, eventually prompting
Congressional action to regulate grazing on the
public domain and to create experimental ranges
such as the Santa Rita (courtesy of Arizona Historical Society, Tucson, AHS #45866).
USDA Forest Service Proceedings RMRS-P-30. 2003
Recognizing History in Range Ecology: 100 Years of Science and Management...
for experiments funded by these monies. It was not until 5
years later, however, that the Federal government took the
decisive step of reserving land from the public domain
specifically for range research. President McKinley signed
the order withdrawing four sections southeast of Tucson on
October 10, 1900.
David Griffiths of the Arizona Agricultural Experiment
Station had spent “the greater part of a week” scouting the
Tucson basin for this tract of land (Griffiths 1901: 23). That
it was bisected by the Southern Pacific Railroad was an
advantage in his eyes, because it meant that one side of his
research plot was already fenced. He enclosed 52 acres,
divided it into 60 plots, and began a series of experiments.
But the “small inclosure” [sic], as it came to be called, soon
showed serious limitations. Even including the unfenced
portion, it was too small and too uniform to represent the
varied rangelands of the region. It was also lower and drier
than the prime grasslands south and east of the Tucson
basin. Griffiths tried to assert that the area was “a typical
mesa region in every respect” (1901: 24), but it contained
more creosote and cacti than perennial grasses and little
topographical, climatic, or edaphic variation. “The production of forage is so small here, at best, that one is obliged to
measure his pasture by square miles rather than by acres,”
he noted, “and the operations in range improvement must
be on a correspondingly large scale” (1901: 29). Even if
Griffiths’ experiments in establishing forage plants had
succeeded, the need for a larger research range would have
remained.
In 1902, Alfred Potter—who would shortly become Gifford
Pinchot’s first Chief of Grazing—drafted a report for the
proposed Santa Rita Forest Reserve, from which the experimental range would subsequently be carved (Potter
1902). In its earliest conception, the reserve was to extend
from the Santa Cruz River east to Cienega Creek, and from
the Southern Pacific Railroad south to Sonoita Creek, an
area of 592 square miles or 379,000 acres. Potter acknowledged that only 45,000 acres of this area was “well forested,” and that nearly four-fifths of it was mesa and
foothills land. Most of the lower elevation, nontimber land
was eventually excluded from the reserve, but on the
northwest flank of the Santa Rita Mountains parts of four
townships were withdrawn, giving birth to the Santa Rita
Experimental Range; Griffiths termed it “the large inclosure” [sic]. President Theodore Roosevelt signed the proclamation on April 11, 1902. The boundaries expanded
under subsequent executive orders, by Roosevelt in 1907,
Taft in 1910, and by Coolidge twice, in 1925 and 1927.
Taft’s order also recognized the Santa Rita as distinct from
the adjacent Forest Reserve, which had been consolidated
into the Coronado National Forest 2 years before; in consequence, title to the experimental range remained with the
Interior Department, rather than transferring to the Department of Agriculture. Ultimately, the Santa Rita encompassed over 53,000 acres, or more than 1,000 times the
size of Griffiths’ first enclosure (which was converted to
military uses in 1925 and today is part of Davis-Monthan
Air Force Base).
Two of Potter’s observations about the Santa Rita Forest
Reserve are worth noting here. First, he wrote that before
the 1891 to 1893 drought, the area had “carried fully 25,000
head of cattle and horses and 5,000 sheep,” and that as of
USDA Forest Service Proceedings RMRS-P-30. 2003
Sayre
1902 these numbers had dropped to “between 7,000 and
8,000 cattle, from 1,000 to 2,000 horses, and about 4,000
sheep.” These figures translate to roughly 44 head per
section before 1891, and 15 to 18 head per section in 1902.
Potter also described a seasonal pattern of movement within
the proposed reserve, with herds concentrating in the mountains in the fall and winter and the foothills in the spring and
summer.
Second, Potter reported that “the mesa lands are all
covered with mesquite, to a certain extent; although over the
greater part of the area the growth is very scattering and
shrubby in character. The only good solid mesquite area is
along close to the river bottom and in the draws coming down
from the mountains.” He described mesquite as “the most
universally useful tree in this section,” providing almost all
the firewood and fence posts used in the vicinity. Since many
wells at this period relied on steam pumps fueled with wood
(Griffiths 1904: 35), it is possible that mesquite harvesting
may have invisibly skewed later perceptions of the area’s
“original” vegetation.
Between March and June 1903, 27.3 miles of fence were
constructed around the experimental range, at a cost of $105
per mile. For the next 12 years, no livestock would graze on
some 49 sections of land, while Griffiths and his successors
studied its recovery. Spanning more than 2,600 ft of elevation, the new reserve encompassed significant gradients of
rainfall, temperature, soils, and vegetation. At the highest,
most productive edge of the reserve, another nine sections of
land were included in the experimental range but were
allowed to remain in the management of settlers already
established there: McCleary, MacBeath, Proctor, and Ruelas.
Their pastures, ranging in size from 194 to 1,695 acres, were
fenced by 1908 and served comparative purposes for the
researchers, suggesting how recovery proceeded under controlled grazing.
That the founding management act of the Santa Rita
Experimental Range was fencing its perimeter is emblematic of circumstances at the time. There had been livestock
in the Santa Cruz Valley for 200 years, and for most of that
time they had not constituted a problem, as far as we know.
Limited transportation and markets, along with notorious
insecurity, had largely isolated the region from outside
sources of livestock, and herd growth had been determined
mainly by local conditions of forage, water, disease, and
predation. Only in the last quarter-century had the cattle
boom flooded the region with livestock from elsewhere,
brought in on foot or by railroad and financed from afar. In
1903, leases and fences were not yet in place to regulate
competition for forage on Federal lands, but there was
finally a political consensus that access to the range had to
be controlled, and that fencing was the only practicable way
to do this. Many early reports implied that fencing, in and of
itself, would cause range conditions to improve; fences went
up on forest lands after 1905, on State Trust lands after
1912, and on the remaining public domain after 1934. The
expanded scale of the livestock industry, from local to international, entailed a contraction of the scale of the individual
herd—from entire valleys or mountain ranges to defined and
fenced pastures. Almost without exception, research on the
Santa Rita would take this geographical innovation for
granted.
3
Sayre
Recognizing History in Range Ecology: 100 Years of Science and Management...
Periods of Santa Rita
Research _____________________
With the range fenced, David Griffiths could begin his
research, initiating the stream of publications by which the
Santa Rita’s scientific production can be appraised. Al
Medina’s (1996) bibliography of Santa Rita research publications lists 452 articles spanning the period 1901 to 1988.
This figure includes 18 undated leaflets and several duplicate entries; excluding these and adding one important
reference omitted by Medina (see below), we have a data set
of 427 articles. If we organize these chronologically, and
depict the results graphically (fig. 2), several periods of
research activity can be identified. This periodization is
intended as a heuristic device only; there have always been
multiple threads of inquiry, administration, and funding in
the fabric of the Santa Rita, and the variable lag between
defining, funding, conducting, and publishing research defies neat temporal separation.
It is immediately clear from the graph that wildlife has
been an important focus of research on the Santa Rita since
the 1920s, increasingly so in recent decades. But it has
rarely constituted a majority of publications, and its very
consistency makes it poorly suited as a means of distinguishing periods of research effort. I defer to Krausman (this
proceedings) to illuminate the place of wildlife research in
the history of the Santa Rita. The history of research on
Lehmann lovegrass is not as long, but otherwise similar to
wildlife—recurrent but minor from the 1940s through the
1980s. I will touch upon it along the way.
1901 to 1931: Institutional Consolidation,
Revegetation, and Carrying Capacity
In the first 30 years of the Santa Rita, only 19 articles
were published, never more than two in any one year. This
was a period of minimal funding and staffing of the range,
while the larger institutional basis for range research was
slowly being consolidated. Major events in this consolidation process included the transfer of the Forest Reserves to
the U.S. Department of Agriculture and the formation of
the Forest Service in 1905; the creation of the Office of
Grazing Studies within the Forest Service in 1910, followed
a year later by the subdivision of forest administration into
25
Other
Fire
Wildlife
Mesquite
Number of publications
20
15
10
5
07
19
10
19
13
19
16
19
19
19
22
19
25
19
28
19
31
19
34
19
37
19
40
19
43
19
46
19
49
19
52
19
55
19
58
19
61
19
64
19
67
19
70
19
73
19
76
19
79
19
82
19
85
19
88
04
19
19
19
01
0
Figure 2—Annual output of publications from Santa Rita research, 1901 to 1988, organized topically. Halfunits reflect publications that expressly covered two topics together (for example, mesquite and fire) (adapted
from Medina 1996).
4
USDA Forest Service Proceedings RMRS-P-30. 2003
Recognizing History in Range Ecology: 100 Years of Science and Management...
regional offices; the transfer of range research outside
National Forests—including the Santa Rita Experimental
Range—from the Bureau of Plant Industry to the Forest
Service in 1915; the transfer of the Office of Grazing
Studies from the Branch of Grazing to the Branch of
Research in 1926; and finally, the creation of regional
forest and range experiment stations under the McSweeneyMcNary Forest Research Act of 1928 (Chapline 1944). It
was this last event that created the Southwestern Forest
and Range Experiment Station (SWFRES), based in Tucson, which brought increased Federal funding for range
research and triggered the rise in Santa Rita publications
from 1932 on.
Two pressing issues dominated the research of this period:
(1) how to restore forage plants decimated by the cattle
boom, and (2) how to measure range resources for management and administration. Griffiths (1901, 1904, 1907) and
J. J. Thornber (1910) tested hundreds of native and nonnative plant species in hopes of finding economical ways of
artificially establishing cover and forage on bare or nearly
bare ground. Most failed altogether, and even those that
showed some success were failures in economical terms.
Building up berms of soil to slow runoff and capture seed was
also attempted at the small enclosure, but the structures
often blew out in floods and did not result in enough grass to
justify the costs. “Much more satisfactory results have thus
far been obtained by husbanding the native vegetation and
grazing well within the capacity of the land to maintain
stock” than by any other methods, Griffiths (1910: 13)
concluded. This recommendation against overgrazing has
been a consistent refrain from Santa Rita researchers ever
since, although far less simple than it appears.
Determining carrying capacities was central to the research of this period because it linked environmental and
ecological factors to political and economic imperatives. It
was of “the utmost importance,” according to Griffiths’
boss, because “This knowledge determines the rental and
sale value of range lands and should also determine the size
of the minimum lease or homestead for range purposes…”
(W. J. Spillman, in the preface to Griffiths 1904). If fencing
and leasing were to work as planned, carrying capacity had
to be a coherent concept that public officials could apply,
measure, and enforce. Furthermore, the capacity of any
given piece of range had to be more or less static, both for
administrative efficiency and so that ranchers and their
financial backers could build leases into their business
plans and credit instruments. Griffiths recognized these
constraints, and he delivered carrying capacity estimates
as best he could, as did Wooton (1916). Following Smith’s
(1899) example from Texas, both were inclined to define
carrying capacity by reference to forage production in poor
(in other words, drought) years. In 1904, Griffiths recommended 37 acres per animal unit (AU) (or about 17 AU per
section) for the Santa Rita generally, and 50 to 100 acres
per AU (about 6 to 13 AU per section) for lower or more
degraded ranges; in 1910 he revised the Santa Rita estimate to 32 AU per section. Wooton concurred with the
latter, higher figure.
Griffiths’ reports contain numerous remarks, however,
that suggest he had doubts about the concept of carrying
capacity when applied “in a region where the seasons, the
altitude, the slope, and the rainfall are so variable” (1904:
USDA Forest Service Proceedings RMRS-P-30. 2003
Sayre
32). Not only did productivity vary across space and time, it
was also “exceedingly difficult to decide which species are
and which are not forage plants,” because, if necessary,
cattle would eat almost everything (1904: 25). Even in the
absence of grazing, the composition of vegetation did not
display stability:
...differences in vegetation, comparing one year with another,
are very striking… In the large field, even with similar
rainfall, there occurs an ascendency [sic] of one plant one year
and another plant another year… So far as known, no one has
ever offered an explanation for these yearly variations of
annual vegetation (1910: 15).
Griffiths also discerned longer term vegetation changes
taking place, specifically an increase in mesquite and other
shrubs, and he attributed these changes to fire suppression,
not grazing. There is no evidence that his doubts diminished
over time; indeed, his 1910 carrying capacity estimates were
even more cautiously expressed than those of 1904. Likewise, his assertion that 3 years of complete rest would
restore degraded Southwestern rangelands “approximately
to their original productivity” (1910: 13) seems forced, because it conflicts with many of his other observations. He
noted, for instance, that 2 consecutive years of good summer
rainfall were needed for significant establishment of perennial grasses—something that occurred only once in his 10
years of research in the area.
Griffiths appears to have arrived in Arizona with few
preconceptions about the desert and no scientific theories to
attack or defend, allowing his curiosity wide latitude. He
conducted surveys of ranchers, traveled and photographed
extensively in the region, and generally let his observations
lead him where they would. In these respects he stands in
sharp contrast to the other major figure of this period,
Frederic Clements, who arrived in Tucson in 1917 to work at
the Carnegie’s Institution’s Desert Lab on Tumamoc Hill
(also founded in 1903). Clements came with a heavy investment in a powerful theory—his own—and a determination
to make it work in the Southwest and, indeed, everywhere.
Clements installed vegetation plots on the Santa Rita
(Bowers 1990: 40), and he also drew on the work of Griffiths
and other Santa Rita researchers for his 1920 book, Plant
Indicators, which included numerous photographs from the
range. The aim of the book was to demonstrate the practical
uses of his famous theory—published as Plant Succession
4 years earlier (Clements 1916)—in managing the rangelands of the American West. The profound influence of
“Clementsian” theory on range science is widely acknowledged to this day (National Research Council 1994; Society for
Range Management 1995). But both Plant Indicators and
Clements’ role in Santa Rita history have virtually disappeared from memory, as evidenced by omission from Medina’s
bibliography. His practical recommendations for managing
livestock in the Southwest have also been largely forgotten,
even though they anticipated many future developments in
semiarid range management. The debates about Clements’
role in range science and ecology have focused on the theory
of plant communities and succession, but I would argue that
the central practical issue was, again, carrying capacity.
Clements suffered from none of Griffiths’ doubts about the
theoretical coherence of carrying capacity, but he defined it
differently and was perhaps more naive than Griffiths about
how it would be used in practice. Specifically, he did not
5
Sayre
construe carrying capacities as static, and apparently he
didn’t see why anyone would.
No other factor produces such rapid and striking changes
in carrying capacity as does rainfall. The difference in the
total yield of the same range in two successive years of
dissimilar rainfall may be greater than 100 per cent, and in
the wet and dry phase of the same cycle it may be even greater
(1920: 292).
Clements believed that some longer “cycle” existed, probably
linked to sun spot activity, which might eventually render
this variability tractable for science and management. But
his practical recommendation was unequivocal:
It is evident that the maximum production can not have a
fixed or average value… A degree of grazing which would be
disastrous in a drought period would fall far short of adequate
utilization during a wet one (1920: 296).
His book exhaustively classified and described Western
rangelands, but he nowhere offered numerical estimates of
acres per animal or animals per section.
It is imperative that the ranchman be prepared to reduce
the pressure upon his range as the dry phase of the climatic
cycle approaches and that he be ready to take full advantage
of the excess carrying capacity of the wet phase. In fact, the
whole system of improvement must be focused upon the
destructive effect of overgrazing in dry years and the possibility of greater utilization and of successful sowing and planting during wet years (1920: 311).
Clements also suggested that carrying capacity was a
function not just of a given range and its condition but also
of management. He criticized both overstocking and stocking year around (1920: 297). Making reference to wild
herbivores such as bison, he linked secondary succession to
long periods of rest following heavy grazing (1920: 307), and
he recommended rotation of grazing pressure to imitate this
natural process (1920: 310). Like Griffiths, Thornber, and
Wooton before him, Clements called strenuously for fencing:
It is immaterial whether control is secured through ownership or leasing, provided it permits fencing. However, leasing
has the indirect advantage that it enables the State to exact
certain conditions as to utilization (1920: 311).
Although Clements’ theory of succession dominated twentieth century range science, as is well known, his practical
recommendations did not dominate actual management.
There is some evidence that southern Arizona ranchers
practiced summer season rest and variable stocking in the
1920s (Sayre 2002), but whether they took their cues from
scientists is unknown—I would guess they did not, in view
of the fact that continuous yearlong grazing became the
norm when ranchers shifted from stocker to cow-calf operations in the 1930s and 1940s. Even among range scientists,
Clements’ theory did not catch on quickly, if we may judge
from the Santa Rita archive. In the minutes of the Forest
Service’s District 3 Grazing Studies Conference of December 1921, for example, there is no mention of the work of
Clements (or Sampson), nor of succession or climax communities. With one exception (Wooton 1916), Clements’
influence does not appear in Santa Rita publications until
the late 1930s.
Although the number of publications from this period was
small, their importance to subsequent research and range
6
Recognizing History in Range Ecology: 100 Years of Science and Management...
administration was great. Griffiths and Clements were
pioneers of range science both at the Santa Rita and for the
nation. That both of them expressed reservations, tacitly or
explicitly, about the central premise of the system of rangeland administration institutionalized over the following
decades sheds new light on current debates about range
ecology and management in the United States and elsewhere (Illius and O’Connor 1999).
1932 to 1945: Growth and the Shrub
Problem
The second period extended from 1932 to the end of World
War II, which imposed severe budget restrictions and brought
publications nearly to zero by 1945. With a newly enlarged
staff, the Southwestern Forest and Range Experiment Station supported more focused studies of particular forage
species such as tanglehead, black grama, blue grama, and
vine mesquite, as well as of noxious or invasive plants,
particularly burroweed. Numerous studies sought more
accurate and efficient methods of measuring vegetation and
utilization—an outgrowth of the Forest Service’s need to
define and enforce carrying capacities. These new methods
were both scientifically rigorous and practical for agencies,
but they did not really address the question of static versus
dynamic carry capacities on Southwestern rangelands.
Revegetation remained a major focus, but with more attention on underlying ecological factors such as litter cover and
soil moisture. Research on wildlife expanded as well to
include kangaroo and pack rats, wood rats, quail, jackrabbits, and rattlesnakes. Finally, there was a more specialized
attention to practical management issues as viewed from
the perspective of private ranchers. Matt Culley (1937)
produced a detailed study of the economics of one of the
Santa Rita’s cooperating ranches, and he and Kenneth
Parker placed numerous articles in livestock journals on
range and management issues such as poisonous plants,
drought, and proper stocking.
Perhaps the most important study performed during this
period, historically speaking, was one that was not published. “Occurrence of Shrubs on Range Areas in Southeastern Arizona” (Upson and others 1937) was a cooperative survey conducted in 1936 and 1937 by the Southwestern
Forest and Range Experiment Station, the Arizona Agricultural Experiment Station, the Agricultural Adjustment
Administration, and the Bureau of Agricultural Economics. It involved vegetation measurements at 450 sites coupled
with ocular surveys of nearly 12 million acres, resulting in
maps of the dominant vegetation covering all of southeastern Arizona (figs. 3 and 4). Nearly a third of the area was
dominated by grasses, and another quarter by creosote;
cactus and burroweed dominated just over 9 percent each,
and mesquite dominated another 7 percent; wolfberry,
saltbush, and snakeweed dominated the remaining 10
percent. The mesquite, snakeweed, and burroweed areas
were singled out as having expanded in recent memory,
usually at the expense of grasses, and therefore as having
the greatest potential for restoration. Up to this point,
burroweed had received far more attention from Santa Rita
researchers than the other two species, but the survey
found mesquite to be the most widespread of the three,
USDA Forest Service Proceedings RMRS-P-30. 2003
Recognizing History in Range Ecology: 100 Years of Science and Management...
Sayre
ARIZONA
KEY M
AP
AREA SURVEYED
FIG-1
PRINCIPAL PLANT COVER OF RANGE AREAS
IN
SOUTHEASTERNARIZONA
CREOSOTEBUSH BALANCE SNAKEWEED, MESQUITE,
CACTUS, & MINORSPECIES, TRACE OF GRASS
BLACKBRUSH BALANCE CREOSOTEBUSH, MESQUITE,
SNAKEWEED, BURROWEED, & MINORSPECIES, SOMEGRASS
CACTUS BALANCE MIXED COMPOSITIONOF MANY
SPECIES, SOMEGRASS
SALTBUSH
TRACES OF MESQUITE,
CREOSOTEBUSH,AND GRASS
WOODLAND-TIMBER
BROWSE & GRASSBENEATH
MESQUITE BALANCE BURROWEED,
SNAKEWEED, & MINORSPECIES, SOMEGRASS
BURROWEED
BALANCE MESQUITE,
SNAKEWEED, CACTUS, & MINORSPECIES, SOME
GRASS
SNAKEWEED
BALANCE MESQUITE,
BURROWEED, BLACKBRUSH, & MINORSPECIES,
SOMEGRASS
GRASSLAND THREE-QUARTERS GRASS, BALANCE
MESQUITE, BURROWEED, SNAKEWEED, & MINORSPECIES
BOUNDARIES OF CROOK AND CORONA
DO NATIONAL FOREST
Figure 3—Map of “principal plant cover of range areas in southeastern Arizona,” 1936 to 1937. The report containing
this map drew attention to shrub encroachment in areas formerly dominated by grasslands, and helped shift attention
from burroweed to mesquite (Upson and others 1937).
present on more than 9 million acres—three-quarters of
the region. Understanding, explaining and remedying this
shift would be the dominant research priority of Santa Rita
range scientists for decades to come.
As mentioned above, Griffiths had noted the spread of
mesquite nearly 30 years earlier and had attributed it
primarily to fire suppression, not grazing. Curiously, “Occurrence of Shrubs” did not discuss fire at all, aside from a
brief mention under “Artificial Means of Control” of
burroweed (p. 26). The report’s explanation of shrub expansion was that grazing—and only grazing—had shifted the
competitive balance between grasses and shrubs, and that
heavily grazed areas around water sources had provided
sites for establishment and subsequent spread of shrubs into
the surrounding range (p. 12–15). This argument was framed,
moreover, in explicitly Clementsian terms: grasslands “represent, of course, the climax type” for the region, and evidence of former grass dominance in areas of shrubs was
USDA Forest Service Proceedings RMRS-P-30. 2003
taken to indicate that such areas “may also be considered,
ecologically, a climax grassland type” (p. 12). The authors
argued that reducing or eliminating grazing would retard or
prevent shrub encroachment, although they also acknowledged documented cases where this had not worked, suggesting the possibility that “there are other factors than
grazing which favor the spread of shrubs” (p. 24).
Researchers initiated studies of mesquite immediately
following completion of the report, but they did not focus on
adjusting stocking rates. Instead, techniques of killing the
trees outright were tested (Parker 1943). In 1940, a study
was launched in which mesquite and/or burroweed were
killed on 1-acre plots; it was followed in 1945 by another,
which used prisoner-of-war labor to thin mesquite to various densities on 2-acre plots. Also in 1940, the Carnegie
Institution ceased its support of the Desert Lab and turned
its facilities on Tumamoc Hill over to the Forest Service.
The SWFRES had its headquarters there until 1953, when
7
Sayre
Recognizing History in Range Ecology: 100 Years of Science and Management...
FIG.-4
OCCURRENCE OF MESQUITE ON RANGE
AREAS
IN
SOUTHEASTERN
ARIZONA
(NUMBER OF PLANTS ACRE)
PER
LIGHT ST
AND
UNDER30 PLANTS
MEDIUM STAN
D 30 TO50 PLANTS
HEAVY STAN
D
OVER80 PLANTS
BOUN
DARIES OF CROO
K & CORONA
DO NATIONA
L
FORESTS
SCALE IN MILES
Figure 4—Map of “occurrence of mesquite on range areas in southeastern Arizona,” 1936 to 1937. The
ubiquity of mesquite, present on roughly 75 percent of the region, led to intense research efforts on the
Santa Rita for four decades, and especially from 1946 to 1966. At the time this map was made, however,
mesquites exceeded 30 plants per acre on only 15 percent of the region; heavy stands (>80 plants per
acre) were confined almost entirely to major drainages (Upson and others 1937).
it was merged into the Rocky Mountain Forest and Range
Experiment Station (RMFRES), headquartered in Fort
Collins, CO.
1946 to 1965: Age of Mesquite
The pace of research rebounded quickly with the end of the
war, and the focus turned decisively to mesquite. The overall
goal was the same as in Griffiths’ day: restoration of perennial forage grasses. But now shrubs, rather than just bare
ground, stood in the way. The postwar period was a prosperous one for both ranchers and agencies, and practices previously deemed uneconomical might now pencil out. Beginning in the late 1940s, the Hope-Flannagan Research and
Marketing Act made funds available for research on noxious
range plants.
8
In 1948, with cooperative agreements up for renewal on
both the Santa Rita and the Jornada Experimental Range,
Kenneth Parker composed “An analysis of range problems
in the Southwest,” another internal document. He cited the
1937 shrub survey in support of the claim that “mesquite
constitutes a problem on some 8 million acres” in southern
Arizona (p. 57)—this appears to have been an exaggeration, as the survey had found medium and high densities of
mesquite on less than 2 million acres (fig. 4). Parker
rejected the earlier study’s Clementsian expectation that
reduced grazing would reverse the trend toward shrub
domination, however. Meter-square quadrats going back to
1916 indicated no consistent relation of vegetation with
either climate or grazing pressure (p. 73); herbage productivity had declined substantially, even with steadily reduced stocking rates (p. 77–79). Parker concluded that “no
degree of moderation in grazing use will eliminate these
USDA Forest Service Proceedings RMRS-P-30. 2003
Recognizing History in Range Ecology: 100 Years of Science and Management...
low value plants. The meaning, in unmistakable terms, is
that if we are to continue grazing use by domestic livestock
some positive, drastic treatment which will eliminate these
plants will be necessary to achieve conservation of the
grazing resource” (p. 71). Because shrub encroachment
also threatened watershed function and, therefore, agricultural and municipal water supplies, Parker argued that
“[t]he future welfare of the Southwest is dependent on how
well and in what manner the range resource is used” (p. 7).
During the severe drought of the early 1950s, a sense of
emergency pervaded the ranching industry, and the “war
on mesquite” played well in local newspapers (Sayre 2002).
In all but 3 years from 1947 through 1965, no less than
one-third and often as many as two-thirds of Santa Rita
publications focused on velvet mesquite (49 out of 109 papers
altogether). Studies ranged from basic questions of life
history and reproduction, to demographic analyses, to effects on soils and competition. Herbicidal approaches to
mesquite control using diesel oil or chemicals were increasingly prevalent in the publications of this period. As in
Griffiths’ day, efforts were launched to find (or create by
hybridization) a perennial grass capable of establishment on
degraded semiarid rangelands, and this time several were
found among South African lovegrasses, although the full
implications of this success would not be evident until the
late 1960s. Work on small mammals also continued, and
whereas many earlier rodent studies had emphasized negative impacts on grasses, now some researchers focused on
rodents’ role in helping to propagate shrubs. Other wildlife
research in this period included studies of javelina, cactus
wren, Gambel quail, and deer.
This body of research has been of major and lasting
significance to scientific understanding of semiarid grassshrub rangelands, even though it fell short of its own
goals for practical management. From a theoretical perspective, the decisive turn was from the Clementsianism
of the 1937 shrub report to Parker’s observation in 1948
that reducing or eliminating grazing would not by itself
cause a reassertion of grasses. This opened up research
questions that extended well beyond issues of livestock
production, laying the foundation for subsequent investigations into water cycling and erosion, the spatial and
temporal distribution of moisture and nutrients as it
affects plant growth and competition (fig. 5), and the role
of small mammals and invertebrates in semiarid ecosystem processes. These issues would emerge to dominate
Santa Rita research in the following period. At the time,
however, the concerns of range managers still focused
primarily on producing livestock, and from this perspective the research fell somewhat short. The methods
developed for controlling mesquite were effective only if
the larger economics of ranching were very favorable—
cheap diesel and high prices for calves—and only on
fairly short time scales of 10 to 20 years, as mesquite
steadily reclaimed treated lands. Although vast acreages would be cleared over the 30 years from 1950 to
1980, the goal of restoring native perennial grass domination once again proved elusive.
Today, we think we know the reason for this shortcoming:
the near-total absence of fire from Southwestern semidesert grasslands. Fire was likewise missing from most
Santa Rita research and publications of the period. In his
USDA Forest Service Proceedings RMRS-P-30. 2003
Sayre
Figure 5—The temporal distribution of effective
rainfall, 1916 to 1948. Shaded areas on the graph
represent periods when rain fell on successive days
or were preceded or followed by storms of 0.4 inch
or greater during the summer growing season. The
graph reflects growing understanding of the ecology
of major forage plants on the Santa Rita, which were
predominantly C4-pathway perennial grasses limited by the distribution of moisture in space and time,
rather than by gross annual or seasonal rainfall
(USDA 1952).
1948 internal analysis, Parker alluded to Griffiths’ comments on fire suppression, but he did not elaborate on them
or recommend research on the subject. Similarly, “The
Santa Rita Experimental Range” booklet of 1952 (USDA
1952) devoted one-sixth of its text and numerous photos to
noxious plant control, without a single mention of fire.
Somewhat of a maverick, Robert Humphrey—who had originally hired on with the Desert Lab—published numerous
papers making the case that fire suppression was the fundamental cause of woody plant encroachment, and that restoring fire could economically control the problem. But his
argument was based more in natural history than in experimentation, and the idea did not go far, producing only two
Masters theses, one technical bulletin, and two peer-reviewed publications other than Humphrey’s own articles
during this period. Ranchers, agencies, and the general
public were all accustomed to vigorous fire suppression, and
the real-world risks were obviously high. Moreover, in Parker
and Martin’s (1952: 14) words, “[t]he effect of fire or lack of
9
Sayre
fire on the occurrence of mesquite stands is a moot question.”
Whether from grazing, drought, shrub encroachment, or a
combination of the three, much of the region’s rangelands
simply didn’t have enough herbaceous fuels to carry a fire.
Lehmann lovegrass had the potential to change this, however, and Humphrey stood alone on this subject, too, calling
attention to the possible downsides of Eragrostis
lehmanniana a decade before anyone else (Humphrey 1959).
The attention placed on chemical and mechanical mesquite control cast a long shadow, obscuring less exciting
topics such as grazing management. Recommended practices were not much changed from earlier periods: stock
conservatively, distribute grazing pressure evenly, defer or
minimize grazing pressure during the summer growing
season (USDA 1952). Echoing Clements, the 1952 Santa
Rita booklet documented the wide variability of rainfall and
forage production, and it cautioned “that the practice of
building up numbers in the occasional good years removes
the only chance that the range might have to improve”
following drought (USDA 1952: 14). But it sought, nevertheless, to establish atemporal guidelines for stocking rates and
utilization (fig. 6).
Figure 6—Graph of forage yield and rainfall on an
annual basis, 1925 to 1933. Dashed horizontal lines
indicate average rainfall (top) and average forage
production (bottom); the solid horizontal line signifies the—“proper basis of use,” defined as roughly
20 percent below average forage production. Intended to prevent overgrazing during recurrent dry
years, this guideline nevertheless perpetuated the
conception of carrying capacity as a static attribute
of Southwestern rangelands (USDA 1952).
10
Recognizing History in Range Ecology: 100 Years of Science and Management...
Meanwhile, postwar prosperity allowed greater capitalization of many ranch operations, and mesquite control was
only one of a long list of investments ranchers were making:
in improved breeding, more fencing and water development,
and new technologies for handling cattle (for example, holding corrals, squeeze chutes, calf tables, pickup trucks, and
trailers). The “old ways of doing things on the range… were
romantic, and led to a simpler and more friendly way of life,”
according to the booklet. “However, they cannot compete
with the modern way of doing business,” which involved
replacing labor costs with fixed costs (USDA 1952: 9). By
this time, cow-calf operations were the norm, and the
booklet recommended dividing one’s herd into groups of 50
to 100 animals, each group with its own fenced pasture, to
allow closer supervision and control of breeding. This
amounted to continuous yearlong grazing, which became
the norm in the region during this period.
1966 to 1988: Ecology and the Santa Rita
Grazing System
Research on mesquite continued through the 1970s, but
its dominance waned. The period 1966 to 1988 was the most
prolific in the Santa Rita’s history, and the proportion of
publications devoted to mesquite declined to only 10 percent of the total, compared to 45 percent in the previous
period. A wide array of new research foci emerged, reflecting new interests and methods both in range science and in
ecology more generally. In the late 1960s, animal scientists
used fistulated steers to study cattle diets, nutrition, and
weight gain, and the idea of frequent, automated weighing
of livestock was pursued. In the early 1970s, the International Biological Program’s Desert Biome project produced
a small mountain of research on soil nutrient flows, soil
moisture, termites, and ants. Other research also looked
below the surface of the ground to examine root systems of
grasses in grazed and protected sites, competition among
plants for soil moisture, variations in soil temperature,
factors affecting runoff and infiltration, fungi, and the
penetration and breakdown of various chemicals, especially insecticides. Wildlife research picked up considerably in the late 1970s, comprising more than a third (22 of
63) of all publications from 1976 to 1981. Different research
activities fed off one another, symbiotically or parasitically
depending on your perspective: mesquite removal for range
restoration experiments raised the question of wildlife
habitat effects; the discovery that termites consume large
quantities of biomass provoked attempts to control them,
just as had happened in earlier decades with rabbits and
rodents.
The problem of mesquite had not gone away. Rather,
confidence and funding had dissipated relative to other
interests. Herbicidal methods had largely failed, and from
the oil crisis of the early 1970s on, the cost of mechanical
treatment could not be justified given stagnant real returns to livestock and the likelihood that retreatment
would be necessary down the road. Where large-scale
mesquite clearing continued, it was underwritten by real
estate appreciation and other nonranching investments,
and it was motivated at least partially by tax policies that
incentivized losses (Sayre 2002). Meanwhile, opposition to
USDA Forest Service Proceedings RMRS-P-30. 2003
Recognizing History in Range Ecology: 100 Years of Science and Management...
mesquite control, especially using chemicals, emerged
among nonranchers as part of the larger social concern for
the environment.
The turn to a broader ecological orientation was reflected in range science research by Clark Martin’s work
on grazing systems. Building on earlier Santa Rita findings about the timing of forage growth, Martin had initiated studies of various grazing/rest schedules on small
plots beginning in 1957, and in the early 1970s he concluded that spring-summer rest 2 years out of 3 produced
significant improvement in perennial grasses compared
to continuous yearlong grazing (Martin 1973). He anticipated that these improvements would be concentrated in
areas of poor range condition (Martin 1978), a prediction
later confirmed in a 10-year study (Martin and Severson
1988). Perhaps the most intriguing discovery of Martin’s
research, and of related work by Dwight Cable (1971,
1975), was that grazing and drought had interactive,
lagged effects extending over 24 to 36 months: significant
improvement resulted from 2 successive years of good
summer rains, and grazing could retard recovery during
the first postdrought summer. These findings echoed the
views of Griffiths, Clements, and the 1952 Santa Rita
booklet, supporting them with hard data.
Rest-rotation grazing was not new, of course. It had
antecedents in the work of Clements, among others, and the
idea of deferring grazing until late in or after the growing
season had been promoted in the 1910s by Jardine and Hurtt
(1917) and Sampson (1914). What was new, it appears, was
a commitment within the Forest Service to encourage the
implementation of rotational systems on allotments throughout the Southwest. Hormay and Talbot (1961) had revived
and systematized rest-rotation in the early 1960s, pointing
out that under yearlong systems selective grazing would
disproportionately impact palatable species, even at conservative stocking rates. Only periodic rest could prevent this,
and fairly heavy grazing could be beneficial if it reduced
selectivity. Hormay and Talbot even claimed that “grazing is
eliminated as an environmental factor under rest-rotation
grazing” (p. 40). Whether true or not, their claim completed
a paradoxical evolution in range science. The discipline had
long embraced Clements’ theory of succession while neglecting his practical management ideas. Now it embraced one of
his management ideas (without crediting him), and used it
to renounce one of the central tenets associated with his
theory: the primacy of grazing in determining vegetation.
Cable’s (1975) research indirectly supported this view by
documenting the overriding importance of summer rainfall.
1988 to Present: Land Swap and
Reorientation
Medina’s bibliography extends only to 1988, and without
knowing his methods and criteria I am reluctant to attempt
to update it. The date would be an arbitrary endpoint for
historical analysis, except that it was also a pivotal year in
the administration and ownership of the Santa Rita Experimental Range. Funding for Santa Rita range science research had been stagnant or declining since 1975, when the
Tucson-based Southwestern Station of the Rocky Mountain
Forest and Range Experiment Station had been merged into
USDA Forest Service Proceedings RMRS-P-30. 2003
Sayre
the Experiment Station at Tempe. Relative to the Southwest’s
booming urban and suburban economic sectors, livestock
grazing had begun to appear less significant, and by the late
1980s the Santa Rita was in danger of becoming an expensive anachronism. That the title to the range still resided in
the Interior Department—a fact that many people had
overlooked, it seems—now became significant. It meant that
the Rocky Mountain Station, and the USDA as a whole,
could simply walk away from the range in response to
shifting priorities and limited budgetary resources. This
would leave it in the hands of the BLM, inheritor of all
undisposed General Land Office holdings. But the BLM did
not have resources or reason to manage an experimental
range either.
Resolution came rather hastily and from an unexpected
direction (Sayre 2002). Some 50 miles southwest of the
range, in the Altar Valley, another branch of the Interior
Department faced a difficulty. The Fish and Wildlife Service
(FWS) had purchased the Buenos Aires Ranch in 1985,
mainly for the purpose of restoring the endangered masked
bobwhite quail. The ranch included leasehold to nearly
90,000 acres of State Trust lands, intermixed with 21,000
acres of deeded land. The FWS had removed all livestock
from the new Buenos Aires National Wildlife Refuge, and it
had no intention of grazing there. This meant, by policy, that
the State Land Department had to reclassify the leases as
commercial and charge the FWS commercial rates: 10 percent of fair market value of the land, or more than half a
million dollars a year. Several small land exchanges were
formulated, which would have enabled the refuge to consolidate its ownership of the prime masked bobwhite habitat.
But the vast majority of the Buenos Aires lease lands would
have remained subject to reclassification, or to reassignment to livestock operators.
Following a change in the governor’s office in 1987, the
Land Department began to press its case and the Buenos
Aires lease fees started to increase, forcing regional FWS
officials to scramble to cover the payments. Early in 1988,
Regional Director Michael Spear and Arizona BLM Director Dean Bibles came up with a solution, which passed into
law with the Idaho-Arizona Conservation Act that November. Nearly two and a half years later, in April 1991, the
transaction was executed: The Interior Department got the
Buenos Aires lease lands, and the State Land Department
took possession of the Santa Rita. Under a special designation passed by the Arizona legislature, the experimental
range was rededicated to research and education. It was
also assigned as its beneficiary the University of Arizona,
which administers the range and collects lease payments
directly from cooperating graziers. In this way, the Santa
Rita conforms to the constitutional mandate of the State
Trust to generate revenue for beneficiaries, but it is outside
of the ordinary policies and procedures of the Land Department. The designation remains in place indefinitely, until
and unless superseded by legislative action (Mitch
McClaran, personal communication).
Under its new ownership, the Santa Rita has continued to
host research projects and to work with its cooperating
grazing lessees. The larger social, economic, political, and
scientific context has shifted dramatically since 1903, however, and the orientation of research on the experimental
range is changing to reflect new interests, opportunities,
11
Sayre
and constraints. Issues of forage and livestock production
are receding relative to those of climate change, ecological
restoration, watersheds and wildlife. I will return to this
reorientation in a moment, after considering the second
question with which I began.
Effects on the Range ___________
What difference did Santa Rita research make on Southwestern rangelands? The question is surprisingly difficult to
answer.
Many innovations developed or recommended by Santa
Rita researchers have been widely adopted: the installation
of water sources every 2 to 3 miles across the range and the
careful placement of mineral supplements to distribute
grazing pressure evenly; the use of improved breeds and
livestock handling techniques; various methods of brush
control and revegetation with grasses; the construction of
interior fences to control both breeding and grazing; rigorous
culling of underperforming animals; and myriad variations
on rotational grazing. Exactly where and when these practices have been implemented, however, and to what effect on
range conditions, are difficult to determine. Grazing impacts
have probably been made more homogeneous and less severe over the landscape, with differential effects depending
on the scale and organism of concern. Lehmann lovegrass is
established in most of the areas suited to it; whether it is
choking out native grasses or otherwise causing harm is still
a matter of debate, but it has unquestionably succeeded in
reducing erosion compared to former conditions of shrub
dominance. Many ranchers now understand the historical
role of fire in these landscapes, and some are working
diligently to restore it; how widely this will succeed, it is too
early to tell.
One core message—avoid overgrazing—has been a constant of Santa Rita management recommendations, along
with the goal of restoring perennial, warm-season grasses.
Beginning early in the twentieth century, these came together in policies focused on proper stocking of National
Forest allotments; later, a similar approach was applied to
BLM and State lands. Clearly, proper stocking was, and
remains, central to good range management. But what did
it mean in practice, and what role did research play in actual
stocking decisions?
It is generally known that forage production and stocking
rates, as well as carrying capacity figures, have declined
significantly in the past 125 years. The stocking rates
recommended by Clark Martin in 1975, for example, ranged
from less than 4 to 18 to 25 AU per section, depending on
condition and elevation (Martin 1975: 10); these are all lower
than the rates recommended by Griffiths in 1910 and Wooton
in 1916, and less than half of actual rates described by Potter
for the area pre-1891. Actual stocking of the Altar Valley
before 1920 was three to fives times greater than at present
(Sayre 2000). As with the West as a whole, assessments of
regional range conditions have been sporadic and hampered
by inconsistent or disputed methodologies (National Research Council 1994).
Excessive grazing is usually viewed as the major cause of
these declines. The agencies were expected to enforce stocking rates, but on the expansive range compliance had to be
12
Recognizing History in Range Ecology: 100 Years of Science and Management...
largely voluntary, and there is evidence that overstocking
was widespread in the past. Using sales data and interviews
for 160 ranches that changed ownership from 1957 to 1963,
Martin and Jefferies (1966) found that actual stocking of
BLM and State Trust allotments was, on average, twice the
official rates. Stocking decisions on State lands were largely
at ranchers’ discretion until the early 1980s, and it seems
that every old timer has stories to tell of permittees who
chronically overstocked their Federal or State allotments.
For obvious reasons, however, more comprehensive data on
the extent and severity of overstocking are extremely difficult to find.
Any assertion of causality between overstocking and
range depletion must be qualified, however. A recent analysis of regional vegetation change argues that the drought of
the 1890s might well have resulted in widespread arroyo
formation even if unaccompanied by overgrazing (Turner
and others 2003). Likewise, the drought of the 1950s
appears to have pushed some Southwestern rangelands—
with and without livestock grazing—across thresholds from
which a return to climax has not occurred (Herbel and
Gibbons 1996). Studies such as these suggest that grazing
impacts may have been significant during periods of severe
drought and much weaker, or even nil, during wetter
periods. When summer rains were good, conditions could
improve in the direction of the “climax” of perennial bunchgrass dominance, even under rates of stocking that we now
characterize as excessive. This occurred, for example, in
the upper end of the Altar Valley in the 1930s (Sayre 2002).
Depletion appears to have been concentrated in drought
periods, when herbaceous vegetation could decline significantly even without livestock present.
It can plausibly be argued—although not proven—that
the changes in vegetation observed during the twentieth
century would have occurred even if actual stocking had
adhered to official carrying capacity estimates. Very likely,
those estimates were unnecessarily restrictive during wet
years and too permissive during severe droughts. In spite of
the great natural variability in forage production, ranchers
had obvious economic incentives to maintain their herds,
even at the risk of overgrazing. “The general practice of
stockmen takes no account of the great variation in yield
between the dry and wet phases,” complained Clements
(1920: 297); this sentiment recurs in reports and bulletins
throughout much of the century. Of course, actual stocking
was never completely static, and carrying capacity estimates continued to be debated, studied, and revised throughout the century. But the expectation that some correct
number of livestock should exist for each allotment, independent of time, was a misconception perfectly suited to strain
relations between agencies and lessees. How could the
agencies ever demand reductions below official capacities,
even in severe drought, if the figures were supposed to
account for poor years? Conversely, how could lessees take
official capacities seriously in wet periods, when forage was
many times greater than permitted numbers of animals
could consume? Range scientists generated carrying capacity estimates that aspired to be independent of fluctuating
rainfall, and economic and political constraints compelled
ranchers and agencies to interpret proper stocking in terms
of static carrying capacities—Griffiths’ muted doubts and
Clements’ explicit admonitions notwithstanding.
USDA Forest Service Proceedings RMRS-P-30. 2003
Recognizing History in Range Ecology: 100 Years of Science and Management...
In summary, the effects of Santa Rita research on
regional rangelands are uncertain. Many management
practices have been adopted, although we do not know
how directly to attribute adoption to research findings. In
some cases, such as the shift from stocker to cow-calf
operations, the science may have reflected, rather than
prompted, the actions of producers responding to market
incentives. Santa Rita research did provide a relatively
independent and objective point of reference for agencies
and ranchers as they endeavored to control the number of
livestock grazing on the region’s Federal and State lands.
This appears to have worked reasonably well provided
that moisture was close to normal—although the norm
may itself have been little more than a statistical artifact.
Wet periods probably undermined ranchers’ respect for
agency guidelines (and perhaps the science behind them
as well); dry periods probably undermined agencies’ confidence in ranchers’ judgment and intentions.
Whether observed changes in vegetation are reversible
depends on whether twentieth century erosion has permanently altered the capacity of a given site to support the
earlier vegetation (Turner and others 2003: 261). Where
the answer is yes, overgrazing may have been responsible,
and the threshold was probably crossed during a major
drought. The static conception of carrying capacity—which
Southwestern range scientists did not expressly denounce
until the 1960s (Paulsen and Ares 1961), and which in
practice remains pervasive to this day—may in turn be
viewed as a contributing factor. In view of the writings of
Griffiths and Clements, however, blame should fall not so
much on the science produced from the Santa Rita and
other experimental ranges as on the translation of research
findings into policy and administration. Had Clements’
dynamic notion of carrying capacity been more widely
embraced, it is possible that the shortcomings of his theory
would not be so obvious today: Agencies and ranchers
might have adjusted stocking rates more aggressively, and
the lasting damage of heavy grazing during drought might
have been avoided. Then again, highly variable carrying
capacities might have made Clements’ theory economically
and administratively impractical and precluded its adoption in the first place. Ironically, Clements himself feared
that his theory might serve as “an excuse for overgrazing”
(1920: 310), but whether any ranchers or agency officials
rationalized heavy stocking in this way is unknown.
Conclusions ___________________
The decision to create the Santa Rita Experimental Range
in 1903 rested on at least two interlocking premises. The
first was that it was biogeographically representative of a
large swath of Southwestern rangelands. Within its boundaries could be found conditions of vegetation, topography,
soils, and climate similar to those of some 20 million acres in
Arizona, New Mexico, and Texas (USDA 1952; fig. 7). The
second was that it was a representative management unit,
similar in size to the larger ranches that dominated the
region. Both premises reflected the judgment that the highest economic use of Southwestern rangelands was grazing,
such that research aimed at the needs of ranchers and range
managers could benefit the entire area. A century later this
USDA Forest Service Proceedings RMRS-P-30. 2003
Sayre
judgment no longer holds, and both premises therefore
warrant reconsideration.
The highest uses of rangelands today, economically speaking, are housing development and recreation. Livestock
grazing in and of itself is relatively insignificant from this
perspective, although in combination with other demands—
for open space, wildlife habitat, and watershed function, for
example—the overall value of ranching remains high. Without getting into whether social demands on rangelands
complement or compete with one another, one can safely say
that the “highest and best” use is no longer uniform. Rather,
it varies depending on factors such as proximity to urban
areas, transportation corridors, or recreational hotspots; the
distribution of wildlife species and their habitats; amenity
values such as scenery and fine weather; and the threats
posed by wildfire, floods, and drought to urban and exurban
settlements.
The landscape is further differentiated by the history of
management. Under equilibrial assumptions this was a
secondary matter because the essential features of the range
were fixed by soils and climate and would reassert themselves if given a chance. In theory, once scientists figured out
how things worked on the experimental range, their knowledge could be taken and applied elsewhere. Now things don’t
look so simple, because we understand—at least in theory—
that discrete events or combinations of events may have
shifted conditions in different ways at different places or
times. Some drainages are cut by arroyos, while others are
not. In some valleys Lehmann lovegrass was planted on
large areas and has spread, while in others it is limited to
roadways or absent altogether. Fields cleared for crops in the
early 1900s still show the effects, decades after abandonment. In some places landscape-scale fires have happened in
living memory, although in most they have not. All these
factors are superimposed on the natural variability of rainfall across space and time as well as the complex patterns of
slope, aspect, soils, and vegetation.
There is still a near consensus that native perennial
grasslands are the most desirable state for the region’s
semiarid rangelands, but the goals of restoration are no
longer rooted in livestock production nor measurable in
terms of carrying capacity. Consequently, how to achieve
restoration, and at what cost, are far from clear. New goals
include wildlife conservation, watershed function, open space
for recreation or for scenery, and ecological restoration. Most
of these generate revenues only indirectly, if at all, and they
are often pursued in the absence of long-term, site-specific
data. Where were various wildlife species present at what
points in the past? How many livestock did each watershed
support during the drought of the 1950s? Which arroyos
have grown in recent decades, which have aggraded, and
what factors are responsible? In summary, a map of the
areas to which knowledge from the Santa Rita might be
applied today would look quite different from the one shown
in figure 7.
The second premise is still true, but less universally so,
and its significance is different from before. Fifty-six thousand acres remains a good size for addressing practical
management problems on Southwestern ranches and on
ranches converted to preserves (if not ranches that have
subdivided). The nature of those problems has changed in
fundamental ways, however, keyed to both spatial and
13
Sayre
Recognizing History in Range Ecology: 100 Years of Science and Management...
Figure 7—Map depicting the geographical areas deemed comparable to the Santa Rita Experimental
Range, 1952. Although based primarily on biogeographical criteria, this judgment of the range of
applicability of Santa Rita research also contained social and economic assumptions, many of which must
be reconsidered in light of dramatic changes in the region’s economy and demography (USDA 1952).
temporal scale. Although the Santa Rita is large, the vast
majority of experiments conducted there have been relatively small (<100 acres, certainly); this reflected both practical constraints and the overriding interest in maximizing
forage production and optimizing utilization. It was generally assumed that findings would extrapolate to larger areas
unproblematically. More recent empirical and theoretical
work casts doubt on this assumption, and today scientists
aspire to landscape-scale observations and experiments.
A parallel change has occurred along the temporal axis.
Most experiments have been less than 5 years in duration,
but longer term data sets have had the most enduring value,
even when they did not lead to publications. Perhaps the
most valuable information derived from the Santa Rita in
the past century, given today’s needs and concerns, is the
long-term series of matched photographs. The power of the
photos is greater than just visual—it derives from their
ability to capture change on a temporal scale unavailable to
the mortal eye and impractical for more sophisticated techniques of data collection. With knowledge and concern about
14
climate change growing, data reaching back a century are
increasingly important. That more research has not been
conducted over periods of about 50 years is regrettable, but
it appears that it took that long for us to recognize the need.
Stewarding and sustaining the Santa Rita is essential for
the continuation of past research and for crafting further
long-term studies designed to answer today’s questions.
Finally, it is worth reconsidering the assumption that
knowledge about rangelands must originate from experiments performed in places such as the Santa Rita. In 1903,
few ranchers had more than 30 years’ experience managing
their lands, and it made sense to aspire to teach them what
could be learned by careful scientific investigation. Today,
there is a significant, albeit shrinking, number of ranchers
whose families carry 100 or more years’ experience in
one place. Their history, and their knowledge, ought now
be understood as a storehouse of genuine and valuable
knowledge for the second century of range science in the
Southwest.
USDA Forest Service Proceedings RMRS-P-30. 2003
Recognizing History in Range Ecology: 100 Years of Science and Management...
References ____________________
Bowers, Janice E. 1990. A debt to the future: scientific achievements
of the Desert Laboratory, Tumamoc Hill, Tucson, Arizona. Desert
Plants. 10(1): 9–47.
Cable, Dwight R. 1971. Growth and development of Arizona cottontop (Trichachne californica [Benth] Chase). Botanical Gazette.
132(2): 119–145.
Cable, Dwight R. 1975. Influence of precipitation on perennial grass
production in the semidesert Southwest. Ecology. 56: 981–986.
Chapline, William R.; Campbell, Robert S.; Price, Raymond; Stewart,
George. 1944. The history of Western range research. Agricultural History. 18: 127–143.
Clements, Frederic E. 1916. Plant succession: an analysis of the
development of vegetation. Publ. 242. Washington, DC: Carnegie
Institution of Washington.
Clements, Frederic E. 1920. Plant indicators: the relation of plant
communities to process and practice. Publ. 290. Washington, DC:
Carnegie Institution of Washington.
Culley, Matt J. 1937. An economic study of cattle business on a
Southwestern semidesert range. Circ. 448. Washington, DC: U.S.
Department of Agriculture. 24 p.
Griffiths, David. 1901. Range improvement in Arizona. Bureau of
Plant Industry Bull. 4. Washington, DC: U.S. Department of
Agriculture. 31 p.
Griffiths, David. 1904. Range investigations in Arizona. Bureau of
Plant Industry Bull. 67 (10 plates). Washington, DC: U.S. Department of Agriculture. 62 p.
Griffiths, David. 1907. The reseeding of depleted range and native
pastures. Bureau of Plant Industry Bull. 117 (3 plates). Washington, DC: U.S. Department of Agriculture. 22 p.
Griffiths, David. 1910. A protected stock range in Arizona. Bull. 177.
Washington, DC: U.S. Department of Agriculture. 28 p.
Herbel, Carlton H.; Gibbons, Robert P. 1996. Post-drought vegetation dynamics on arid rangelands of southern New Mexico.
Agricultural Experiment Station Bull. 776. Las Cruces: New
Mexico State University, College of Agriculture and Home
Economics.
Hormay, A. L.; Talbot, M. W. 1961. Rest-rotation grazing: a new
management system for perennial bunchgrass ranges. Production Res. Rep. 51. Washington, DC: U.S. Department of Agriculture, Forest Service. 43 p.
Humphrey, Robert R. 1959. Lehmann’s lovegrass, pros and cons. In:
Your range—its management. Rep. 2. Tucson: University of
Arizona, Agricultural Experiment Station. 28 p.
Illius, A. W.; O’Connor, T. G. 1999. On the relevance of nonequilibrium concepts to arid and semiarid grazing systems. Ecological
Applications. 9(3): 798–813.
Jardine, James T.; Hurtt, L. C. 1917. Increased cattle production on
Southwestern ranges. Bull. 588. Washington, DC: U.S. Department of Agriculture.
Martin, S. Clark. 1973. A grazing system for semidesert ranges
based on responses of grasses to seasonal rest. In: Abstracts of
papers: 26th annual meeting, Society for Range Management;
1973 February; Boise, ID. Denver, CO: Society for Range Management: 21.
Martin, S. Clark. 1975. Ecology and management of Southwestern
semidesert grass-shrub ranges: the status of our knowledge. Res.
Pap. RM-156. Fort Collins, CO: U.S. Department of Agriculture,
Forest Service, Rocky Mountain Forest and Range Experiment
Station. 39 p.
Martin, S. Clark. 1978. The Santa Rita grazing system. In: Hyder,
Donald N., ed. Proceedings of the First International Rangeland
Congress; 1978 August 14–18; Denver, CO. Denver, CO: Society
for Range Management: 573–575.
USDA Forest Service Proceedings RMRS-P-30. 2003
Sayre
Martin, S. Clark; Severson, Keith E. 1988. Vegetation response to
the Santa Rita grazing system. Journal of Range Management.
41(4): 291–295.
Martin, William E.; Jefferies, Gene L. 1966. Relating ranch prices
and grazing permit values to ranch productivity. Journal of Farm
Economics. 48: 233–242.
Medina, Alvin L. 1996. The Santa Rita Experimental Range: annotated bibliography (1903–1988). Gen. Tech. Rep. RM-GTR-276.
Fort Collins, CO: U.S. Department of Agriculture, Forest Service,
Rocky Mountain Forest and Range Experiment Station. 67 p.
National Research Council. 1994. Rangeland health: new methods
to classify, inventory, and monitor rangelands. Washington, DC:
National Academy Press. 180 p.
Parker, Kenneth W. 1943. Control of mesquite on Southwestern
ranges. Leaflet 234. Washington, DC: U.S. Department of Agriculture. 8 p.
Parker, Kenneth W. 1948. An analysis of range problems in the
Southwest. Unpublished report on file in: Archives of the Santa Rita
Experimental Range, University of Arizona, box SRP 60. 130 p.
Parker, Kenneth W.; Martin, S. Clark. 1952. The mesquite problem
on southern Arizona ranges. Circ. 908. Washington, DC: U.S.
Department of Agriculture. 70 p.
Paulsen, Harold A., Jr.; Ares, Fred N. 1961. Trends in carrying
capacity and vegetation on an arid Southwestern range. Journal
of Range Management. 14(2): 78–83.
Potter, Alfred F. 1902. Report of the examination of the proposed
Santa Rita Forest Reserve. Unpublished report on file in: Archives of the Santa Rita Experimental Range, University of
Arizona, box SRP 67. 27 p.
Sampson, Arthur W. 1914. Natural revegetation of range lands
based upon growth requirements and life history of the vegetation. Journal of Agricultural Research. 3(2): 93–147.
Sayre, Nathan F. 2000. Land use history of the Altar Valley. Task
3 of the Altar Valley watershed resource assessment. Tucson, AZ:
Altar Valley Conservation Alliance.
Sayre, Nathan F. 2002. Ranching, endangered species, and urbanization in the Southwest: species of capital. Tucson: University of
Arizona Press.
Smith, Jared G. 1899. Grazing problems in the Southwest and how
to meet them. Division of Agrostology Bull. 16. Washington, DC:
U.S. Department of Agriculture. 47 p.
Society for Range Management. 1995. New concepts for assessment
of rangeland condition. Journal of Range Management. 48(3):
271–282.
Thornber, J. J. 1910. The grazing ranges of Arizona. Bull. 65.
Tucson: University of Arizona, Agricultural Experiment Station.
360 p.
Toumey, J. W. 1891. Arizona range grasses in general; overstocking
the range. Arizona Agricultural Experiment Station Bull. 2.
Tucson: Arizona Agricultural Experiment Station.
Turner, Raymond M.; Webb, Robert H.; Bowers, Janice E.; Hastings,
James Rodney. 2003. The changing mile revisited. Tucson: University of Arizona Press.
Upson, Arthur; Cribbs, W. J.; Stanley, E. B. 1937. Occurrence of
shrubs on range areas in southeastern Arizona. Unpublished
report issued by the Arizona Agricultural Experiment Station.
On file in: Archives of the Santa Rita Experimental Range,
University of Arizona, box SRP 60. 30 p.
U.S. Department of Agriculture. 1952. The Santa Rita Experimental Range. Booklet (revised). Tucson, AZ: U.S. Department of
Agriculture, Forest Service, Southwestern Forest and Range
Experiment Station. 17 p.
Webb, George E. 2002. Science in the American Southwest. Tucson:
University of Arizona Press.
Wooten, E. O. 1916. Carrying capacity of grazing ranges in southern
Arizona. Bull. 367. Washington, DC: U.S. Department of Agriculture. 40 p.
15